The applicant's career goal is to become a productive independent investigator in the area of statistical genetics, particularly in the area of genomic-based prediction of type-2 diabetes (T2D) risk. To meet this goal, the applicant proposes a career development plan that includes hands- on and didactic training in statistical learning as applied to quantitative genetics, categorical and case-control data analysis, informatics as applied to high dimensional genetic data, and the biology and genetics of T2D. A highly accomplished and diverse set of investigators with proven track records of successful mentoring will oversee the applicant's career development. The research component of this project seeks to improve our ability to use genetic information to predict an individual's risk of developing T2D Publically available genetic and phenotypic data from sources such as dbGaP (The database of Phenotypes and Genotypes) will be used to develop and test various models for prediction of T2D risk among three racial/ethnic groups. This project will capitalize on newly developed statistical methods that are able to incorporate information from tens of thousands of genetic markers at once, which represent a major advance over current methods that typically take fewer than 100 markers into account. The aims of the study are: 1) To test the hypothesis, in different populations, that individualized whole-genome prediction of T2D (along with standard covariates of sex, age, and BMI) will offer major improvements over current genetics-based prediction models, and will offer equal or greater accuracy than prediction based on family history; 2) To impute additional genetic markers to determine whether prediction can be improved, and to identify the subset of markers that is most useful in predicting T2D; 3) To develop prediction models for T2D risk given a certain body mass index (BMI), by including as predictors the interaction of BMI and genotypes. This project will greatly enhance our ability to predict an individual's susceptibility to T2D within various populations, leading to earlier and targeted prevention strategies, will increase our understanding of the genetic basis of T2D, and will provide critical training for Dr. Klimentidis' development as an independent scientist. PUBLIC HEALTH RELEVANCE: Genetic factors are known to play a substantial role in the etiology of type-2 diabetes (T2D), which is a growing health challenge facing developed and developing countries. In this application, we propose to use state-of-the-art statistical methods that enable us to account for thousands of genetic variants simultaneously, in order to build and test predictive models for T2D susceptibility among individuals of European, Mexican, and African descent. This project has the potential to greatly improve prediction of T2D susceptibility in several ethnic groups, provide a better understanding of the specific genetic factors that play a role in T2D susceptibility, and ultimately bring us closer to the era of personalized medicine in which improved prediction ability results in earlier patient-oriented prevention and treatment.